Important Note:
This program uses estimation techniques to analyze the data entered. While the results are likely to be similar to the results of an exact calculation process (which requires advanced statistical software), the user should consider the results only as estimations--especially for small data sets or when the results are close to statistical significance (p<.05).
Overview
This part of the Program is designed for comparing the passing rates of each gender and ethnic group on several combined “events” or administrations of various practices, procedures, or tests. It may also be used to complete an overall adverse impact analysis on several jobs or groups of jobs with similar skill sets. It may also be used to compare group passing rates on an overall selection or promotion process for multiple years, although an event-by-event analysis should be the primary comparison in most circumstances [the 1991 Civil Rights Act requires that a “particular employment practice” needs to be identified as the source of adverse impact for a plaintiff to establish a disparate impact case, unless the results are not capable for separation for analysis—see Section 2000e-2(k)(1)(A)(i)].
Processing...
Footnotes:
(1) Selection Rate Difference: The Focal Group's Selection Rate minus the Reference Group's Selection Rate. Negative values show the Focal Group disadvantaged.
(2) Pattern Consistency for Statistical Tests: These statistical tests evaluate whether the "trend" in the passing rate difference between groups is consistent between events. When one group continually maintains a passing rate that is lower than the other group, the result of this test will be close to 1.0. When there are "flip-flops" between events (e.g., when one group's selection rates are below the other group's for four years in a row, then in the fifth year the trend reverses), the value will approach 0. When the result of this test is below .05, a statistically significant "flip-flop" of one or more events has been detected, and the user should consider conducting separate analyses for these events or removing them altogether. The first test uses the "Breslow-Day" (corrected) test which evaluates the appropriateness of the Mantel-Haenszel test in Step 2; the second uses the "Treatment by Strata Interaction," which evaluates the appropriateness of the Minimum Risk test in Step 2 (see below).
(3) Statistical Tests (Mantel-Haenszel): There are two versions of this (two-tail) test (Version 1 and 2). Version 1 is likely to be closest to the "exact" probability value, and only uses a modest (.125) continuity correction. Version 2 uses a conventional correction (.5) and will typically overestimate the probability value (especially with small data sets). These tests assess whether the selection rate difference between two groups (e.g., men vs. women) for all "events" combined is extreme enough to be considered "beyond chance." Values less than .05 (indicated in red) are "statistically significant"; values between .05 and .10 (in orange) are "close" to significance. Both tests use the Cochran version of the Mantel-Haenszel statistic, which weights the events by sample size. A "VALID" sign next to the statistical output indicates that the output can be interpreted because the Pattern Consistency test in Step 1 was not violated; a "WARNING" sign indicates otherwise.
(4) Statistical Test (Minimum Risk Method): This test uses the (two-tail) Minimum Risk test to assess whether the selection rate difference between two groups (e.g., men vs. women) for all "events" combined is extreme enough to be considered "beyond chance." This test is typically a very good estimator to the "exact" probability value, and tends to make very balanced estimations (i.e., it is balanced in making over- and under-estimations). The primary difference between this test and the Mantel-Haenszel is that this test equally weights each event (regardless of the sample sizes of each). Values less than .05 (indicated in red) are "statistically significant"; values between .05 and .10 (in orange) are "close" to significance. A "VALID" sign next to the statistical output indicates that the output can be interpreted because the Pattern Consistency test in Step 1 was not violated; a "WARNING" sign indicates otherwise.
(5) Interpretation of Statistical Test: These outputs describe the degree of the Statistical Test findings. For example, if the output shows the likelihood of the statistical test value is "1 in 20," this means that the difference in passing rates between groups (e.g., men vs. women) is so extreme that the odds of it occurring by chance is only 1 in 20, or about 5%. In other words, this result indicates that chance can be "ruled out" as a reason for this difference. The "Probability as Std. Deviations" describes the probability value (from the Statistical Test) in terms of standard deviations units, which are sometimes easier to interpret than small probability values. A standard deviation of 1.96 corresponds with a probability value of .05, and a likelihood of 1 chance in 20.